NOTES
1148
Of theseIfew examples, two bear on the effect of an adjacent oxygen atom. Burchfield2 found that on pyrolysis of the diacetate of propane-l,2-diol the secondary acetate was eliminated the more rapidly, and 75% of the olefinic product was allyl acetate and 25% propionaldehyde enol acetate. DePuy, King, and Froems2-acetoxy-l-methoxypropane, and d ~ r f pyrolyzed ~ . ~ found an approximately 50-30 mixture of the 1- and 2olefins to be formed. A third ester with a beta oxygen has also been pyrolyzed. This ester is 3-acetoxytetrahydrofuran (I), which is readily prepared from the commercially available 3-hydroxytetrahydrofuran. Olsen5 noted that pyrolysis of this ester gave no 2,3-dihydrofuran (111), but instead gave 3,4-dihydrofuran (11) and furan.
It had previously been shown, however, that 2,3-dihydrofuran decomposes when heated on glass6with the ultimate production of propylene and carbon monoxide. The following schame has been suggested. I11
[kCI-IO]-
co
VOL.
28
Company. Careful fractionation gave a fraction; b.p. 71.572.5", 19 mm. [lit.,' b.p. 64" (2 mm.)]. 3,4-Dihydrofuran (II).-This olefin was prepared by treating 2-butene-l,4-diol ( 2 5 9.) with sulfuric acid (1 g.) in a distilling flask a t 20 mm. The product distilled as soon as formed, and was collected in a Dry Ice trap. Fractionation gave a pure fraction, b.p. 6*5436' (lit.,' b.p. 67"). Pyrolysis.-The pyrolyses were carried out as described previously.8 In a typical run a mixture of 6.61 g. of chlorobenzene and 3.93 g. of I was analyzed by g.p.c. and the relative peak area determined. A portion of the mixture (3.27 g . ) was pyrolyzed a t 450'. At the temperature and flow rate chosen only about 50% conversion was obtained. A low boiling gas, identified as propylene by its g.p.c. retention time was obtained, along with higher boiling products. The higher boiling products were analyzed by g.p.c. using the chlorobenzene as an internal standard. From the decrease in ester-chlorobenzene ratio the amount of ester decomposed could be determined. From the olefinchlorobenzene ratio the amount of 3,4-dihydrofuran was determined. Corrections were made for differences in thermal conductivities between the olefin and chlorobenzene, as determined by the analysis of known mixtures. By this method it was determined that pyrolysis of I gives i 7 f 370 11. Gas Chromatography.-A 1-m. column of THEEIY 1 : 3 on 60/80 firebrick in series with a 1-m. column of THEED on Fluoropak was used a t 125". (7) W.Reppe, A n n . , 696,1 (1935). (8) N,N,N',N'-Tetrakis(2-hydroxyethyl)ethylenediartline.
Half-lives of Some Inert Oxiranes in Acetic Acid-Pyridinium Hydrobromide at 30' i
H2C=CH-CHa
MAXW. MILLER
If this decomposition reaction of the enol ether were taking place during the pyrolysis reaction, a simple product analysis would give misleading results with respect to the direction of elimination. As a consequence we have re-examined this pyrolysis. At 475500' only a single dihydrofuran is formed in the pyrolysis of I, and this olefin is 3,4-dihydrofuran, as shown by its spectral properties and its retention time on gas chromatography. I n addition, two low-boiling products are formed, and these have been identified as propylene and carbon monoxide. We therefore believe that the 2,3-dihydrofuran is destroyed during the pyrolysis. I n order to get an accurate measure of the relative amounts of the two olefins formed, we have made use of gas chromatography with an internal standard. A known mixture of the ester (I) and chlorobenzene was pyrolyzed to about 50% conversion. This conversion was determined accurately by the relative amounts of chlorobenzene and ester in the product. The amount of 3,4-dihydrofuran formed was also determined accurately from the chlorobenzene-olefin ratio in the product. By this method it was determined that pyrolysis of I gives 77 f 3y0 of 11. By difference it is assumed that the other 23y0 is 111. Experimental 3-Acetoxytetahydrofuran (I).-This ester was prepared from 3-hydroxytetrahydrofuran obtained from the Aldrich Chemical (2) P. E. Burchfield, U. S. Patent 2,485,694(1949); Chem. Abstr., 44, 2007 (1950). (3) C. H.DePuy, R. W. King, and D. II. Froetnsdorf. Tetialiedron, 7 , 123 (1959). (4) C. H. DePuy and R. W. Kina, Chem. Rev., 60,433 (1960). ( 5 ) S. Olsen, Acta Chem. Scand., 4, 473 (1950). (6) J. Wilson, J. Am. Chem. Soc., 69, 3002 (1947).
Medical Research Laboratories, Chas. Pfizer and Conqmny, I n c . , Groton, Connecticut Received October 16, 196d
Epoxysuccinic acids have been prepared by a variety of methods. Earlier stereochemical studies involved their interconversion to the corresponding chloromalic and tartaric acids.'-3 Both the trans-racemate and the cis,meso-forms can be prepared from fumaric and maleic acids, respectively, by a sodium tungstatecatalyzed hydrogen peroxide oxidation. The cis,meso-form has been prepared also by the hydrogen peroxide oxidation of benzoquinone5 and by nitric acid oxidation of the macrolide antibiotic, carbomycin.6 To obtain the pure (-)-trans isomer, however, a fermentative preparation is convenient since good yields have been reported from the fermentation of dextrose by Aspergillus fumigatus.7-10 These prior publications, particularly some of the oxidative preparations and the formation of an acid chloride from the cis isomer by prolonged heating with phosphorus pentachloride, l 1 suggested a remarkable (1) R . Kuhn and F. Ebel, Ber., 6 8 , 926 (1925). (2) R.Kuhn and R. Zell, ibid.,69, 2514 (1926). f 3 ) R. Kuhn and T . Wagner-Jauregg. i6id., 61,513 (1928). (4) G.B. Payne and P. H. Williams. J . Org. Chem., 24, 54 (1959). ( 5 ) E. Weita, H. Schobbert, and H. Seibert. Ber., 68, 1166 (1935). (6) R. B. Woodward, Angew. C h e m . , 69, 50 (1957). (7)J. H.Birkinshaw, A. Bracken. and H. Raistrick, Biochem. J . , S9, 70 (1945). (8) J. Moyer, U. 5. Patent 2,674,561(September 8, 1950), t o the Secretary. Department of Agriculture. (9) W. Martin and J. Foster, J . Boeteriot., 70, 405 (195.5). (10) I t is interesting that fumagillin, another Aspergillus lumigatus metabolite, contains two epoxide groups: J. Landquisc, J . C h e m . Soc., 4237 (1956),and D. S. Tarbell, R. RI. Carman, D. D. Chapman, K. R . Huffman, and S . J. McCorkindale, J . A m . Chem. Soc., 82, 1005 (1960). (11) C. Schopf and W. Arnold, Ann., 688, 122 (1947).
APRIL,1963
1149
NOTES
I n an isotopic study of the acid-catalyzed hydrolysis acid stability for the ethylene oxides so substituted. of substituted ethylene oxides it was concluded that We prepared (-)-trans-2,3-epoxysuccinic acid by the cleavage and its rate determining step involved forthe fermentation of crude molasses with Aspergillus mation of a carbonium ion intermediate. 16,17 fumigatus, and obtained a 40% molar yield based on the glucose content. A number of derivatives of the mold metabolite were prepared for structure-activity studies, and the impressive acid stability of this series of compounds seemed to be confirmed. It was possible, for example, to prepare esters by the ordinary Fischer technique, and to prepare the nitrile by heating W the amide with phosphorus oxychloride. Out of curiosity a comparison was made of the halflives of several of the simpler derivatives in acetic acidIf a similar mechanism can be assumed for the pyripyridinium hydrobromide a t 30". Since it was redinium hydrobromide-epoxide transition state, then puted to be an inert oxirane, trans-2,3-dibenzoylethyl- the rate differences observed are consistent with the ene oxidel2 was included in the measurements. An expected relative stabilities of the developing carbonium ions, nitrile and benzoyl groups being less effective I. -H carbonium ion stabilizers than carboxyl, carbomethoxyl, 11. -COOH R or carboxamide groups. / H 111. -COOCHB \ IV. - CONHz c-c, Other derivatives of (- )-trans-2,3-epoxysuccinic bod '"R V. - C N acid as well as some structure-activity relationships will VI. - c o n LJ be described elsewhere.'* attempt was made to compare the compounds with ethylene oxide itself, but the latter proved too reactive for a rate determination under our conditions. Pyridinium hydrobromide was used as a nonvolatile source of hydrobromic acid, but the pyridinium ion also participates importantly in the ring opening. In the cleavage of ethers by halogen halides it increases the rate.13 In the anhydrous reaction of ethylene oxide with pyridine solutions of hydrogen halides an ion-pair complex has been proposed as the transition state, the formation of this complex being rapid, and its decomposition either by rearrangement to halogenohydriii or by attack of solvent molecules a t carbon being the slow step responsible for the rate and the second-order kinetics observed. l 4 , I 5
Experimental19 Fermentative Conversion of Crude C Molasses to ( - )-transEpoxysuccinic Acid.-An improved strain of Aspergillus fumigatus Fres. was used. Initially, the medium was based on those described in earlier publications (e.g.,*), and a medium from this patent is reproduced in Table 11. TABLEI1 &/I.
Glucose monohydrate Beet molasses Corn steep liquor MgSO4.7H2O ("4)804
KH?POI ZnS04.7H20 CaC03 hIethanol
1150.0
20.0 10.0 0.25 .70 .30 ,044 30.0 32.0
I n still fermentations a period of 12 days a t 30" was required.
It was later found poesible to use submerged fermentations for a shorter period with crude C molasses as the sole carbon source. Isolation of Barium Epoxysuccinate.-A 130-gal. batch of fer-
Results of the experiment are shown in Table I. The cleavage rate of epoxysuccinamide, the least reactive of the series, was one thirtieth as great as that of epoxysuccinonitrile and probably less than one thousandth as great as that of ethylene oxide. TABLE I HALF-LIVES OF SOMEIXERT OXIRANES IN ACETIC ACID-PYR~DIKIUM HYDROBROMIDE AT 30" Compound
Ethylene oxide" Epoxysuccinonitrile Dibenzoylethylene oxide Epoxysuccinic acid Methyl epoxysuccinate Epoxysuccinamide Too small to measure by our method,
Half-life t / 2 . min